What does the error " Unable to compile C++ code with dijitso" mean?

1

Hi!

While executing a code, I came across an error message
RuntimeError: Unable to compile C++ code with dijitso

I went through some of the responses to same/similar error messages from the community responses but unfortunately none of them worked.

Here is an MWE

from __future__ import print_function
from fenics import *
import matplotlib.pyplot as plt


from dolfin import *
import meshio
from dolfin import Mesh, XDMFFile, File, MeshValueCollection, cpp



# Optimization options for the form compiler
parameters["form_compiler"]["cpp_optimize"] = True
ffc_options = {"optimize": True, \
               "eliminate_zeros": True, \
               "precompute_basis_const": True, \
               "precompute_ip_const": True}



from mshr import Circle, generate_mesh
from dolfin import Mesh, File, MeshFunction, Point, BoundaryMesh, SubDomain, plot, File
import matplotlib.pyplot as plt
from dolfin import *
from mshr import *


class boundary(SubDomain):
    def inside(self, x, on_boundary):
        return on_boundary


class disk(SubDomain):
    def inside(self, x, on_boundary):
        return x[0] < 0 and x[0] > 0


mesh = generate_mesh(Circle(Point(0, 0), 3), 50)
boundary_markers = MeshFunction("size_t", mesh, mesh.topology().dim()-1, 0)
surface_markers = MeshFunction("size_t", mesh, mesh.topology().dim(), 0)
boundary().mark(boundary_markers, 2)
disk().mark(surface_markers, 1)


ds = Measure('ds', subdomain_data=boundary_markers)
dx = Measure('dx', subdomain_data=surface_markers)
n = FacetNormal(mesh)

W1 = FunctionSpace(mesh, "Lagrange", 1)
n = FacetNormal(mesh)
G , mu = 1, 0.1
u_D=Constant(0.0)

bc = DirichletBC(W1, u_D, boundary_markers, 2)

# Define variational problem
u = TrialFunction(W1)
v = TestFunction(W1)
f = Constant(-G/mu) # f=-G/mu
a = dot(grad(u), grad(v))*dx
L = -f*v*dx

# Compute solution
u = Function(W1)
solve(a == L, u, bc)

file = File("fsp1.pvd");
file << u;

gamma1 = Expression(("(-1/D)*(f0 + f1*(1-exp(-beta*(t**2))))"), degree = 1, D = 0.7, t = u, f0 = 5, f1 = 2, beta = 0.3)

This produces a gigantic error message, a part of which reads

Traceback (most recent call last):
  File "fspex.py", line 70, in <module>
    gamma1 = Expression(("(-1/D)*(f0 + f1*(1-exp(-beta*(t**2))))"), degree = 1, D = 0.7, t = u, f0 = 5, f1 = 2, beta = 0.3)
  File "/usr/lib/petsc/lib/python3/dist-packages/dolfin/function/expression.py", line 400, in __init__
    self._cpp_object = jit.compile_expression(cpp_code, params)
  File "/usr/lib/petsc/lib/python3/dist-packages/dolfin/function/jit.py", line 158, in compile_expression
    expression = compile_class(cpp_data, mpi_comm=mpi_comm)
  File "/usr/lib/petsc/lib/python3/dist-packages/dolfin/jit/jit.py", line 173, in compile_class
    raise RuntimeError("Unable to compile C++ code with dijitso")
RuntimeError: Unable to compile C++ code with dijitso
2

It means that the string here, cannot be interpreted as a C++ string. You need to use the pow( , ) operator to do powers in C++.

1 Like
3

Hi dokken! Thanks for the response.

Your suggestion works perfectly fine for the MWE I created. But with the actual code I am dealing with, the problem still persists after your suggested modification. I am a bit hesitant to post the actual problem since it imports geometries from GMSH which is a quite complicated geometry and cannot be reproduced within FEniCS and also in a previous post, you advised me to create some built-in geometries so that the code can be easily reproduced and I absolutely agree with that. So right now, I have no clue why your suggestion works for the MWE and not in my actual code. If you want I can post the gmsh and the python script.

4

As this relates to an Expression, there should be no reason for it being dependent on the mesh. Please try to simplify the problem by using a built-in mesh, and see if the error persists.

1 Like
5

Ok, I will create an MWE and if there is a problem, I will get back to you.

1 Like
6

Hi dokken,

Here is an MWE with a built-in mesh:

from __future__ import print_function
from fenics import *
import matplotlib.pyplot as plt

from dolfin import *
import meshio
from dolfin import Mesh, XDMFFile, File, MeshValueCollection, cpp

# Optimization options for the form compiler
parameters["form_compiler"]["cpp_optimize"] = True
ffc_options = {"optimize": True, \
               "eliminate_zeros": True, \
               "precompute_basis_const": True, \
               "precompute_ip_const": True}

import numpy as np
import meshio
from mshr import Circle, generate_mesh
from dolfin import Mesh, File, MeshFunction, Point, BoundaryMesh, SubDomain, plot, File
import matplotlib.pyplot as plt
from dolfin import *
from mshr import *

class boundary(SubDomain):
    def inside(self, x, on_boundary):
        return on_boundary

class disk(SubDomain):
    def inside(self, x, on_boundary):
        return True


mesh = generate_mesh(Circle(Point(0, 0), 3), 50)
boundary_markers = MeshFunction("size_t", mesh, mesh.topology().dim()-1, 0)
surface_markers = MeshFunction("size_t", mesh, mesh.topology().dim(), 0)
boundary().mark(boundary_markers, 2)
disk().mark(surface_markers, 1)

ds = Measure('ds', subdomain_data=boundary_markers)
dx = Measure('dx', subdomain_data=surface_markers)
n = FacetNormal(mesh)

W1 = FunctionSpace(mesh, "Lagrange", 1)
n = FacetNormal(mesh)
G , mu = 1, 0.1
u_D=Constant(0.0)

bc = DirichletBC(W1, u_D, boundary_markers, 2)

# Define variational problem
u = TrialFunction(W1)
v = TestFunction(W1)
f = Constant(-G/mu) # f=-G/mu
a = dot(grad(u), grad(v))*dx
L = -f*v*dx

# Compute solution
u = Function(W1)
solve(a == L, u, bc)

# Calculating shear stress S
Ux=0.0
Uy=0.0
Uz=u
U = as_vector((Uy, u))
N = as_matrix([[0.0,0.5],
                 [0.0,0.0]])
Sn = mu/2 * dot(grad(U) + grad(U).T, (N.T)*n)
print(assemble(Sn[1]*ds))
gamma1 = Expression(("(-1/D)*(f0 + f1*(1-exp(-beta*(pow(t,2))))"), degree = 1, D = 0.7, t = Sn[1] , f0 = 5, f1 = 2, beta = 0.3)

And the error reads (I am copying the whole error message):

Traceback (most recent call last):
  File "cae.py", line 70, in <module>
    gamma1 = Expression(("(-1/D)*(f0 + f1*(1-exp(-beta*(pow(t,2))))"), degree = 1, D = 0.7, t = Sn[1] , f0 = 5, f1 = 2, beta = 0.3)
  File "/usr/lib/petsc/lib/python3/dist-packages/dolfin/function/expression.py", line 400, in __init__
    self._cpp_object = jit.compile_expression(cpp_code, params)
  File "/usr/lib/petsc/lib/python3/dist-packages/dolfin/function/jit.py", line 158, in compile_expression
    expression = compile_class(cpp_data, mpi_comm=mpi_comm)
  File "/usr/lib/petsc/lib/python3/dist-packages/dolfin/jit/jit.py", line 173, in compile_class
    raise RuntimeError("Unable to compile C++ code with dijitso")
RuntimeError: Unable to compile C++ code with dijitso

Any suggestion would be greatly appreciated.

7

The parentheses in gamma1 are not balanced. Also, you’re trying to include a ufl form for t=Sn[1] in your Expression where a primitive number, Expression, or Function is expected.

1 Like
8

Thanks Nate! Projection of Sn[1] onto a function space fixed the problem.

1 Like